Correspondence

Cost Savings at the End of Life

N Engl J Med 1994; 331:477-479August 18, 1994

Article

To the Editor:

In their analysis of proposals to control costs at the end of life, such as the use of advance directives and hospice care, Emanuel and Emanuel (Feb. 24 issue)1 somehow conclude that an annual saving of $29.7 billion is less than substantial.

In most people's view, $30 billion in savings is anything but unsubstantial. For example, $30 billion is enough to provide health insurance to uninsured Americans as part of a plan of universal coverage.

These savings can also be critical for people already in the system. On the basis of the Emanuels' estimate, the total savings for the years 1995 through 2000 would be $277 billion -- more than enough to offset the $191 billion in spending cuts proposed for Medicare and Medicaid for that period under President Clinton's Health Security Act (these savings are estimated from projected health expenditures)2.

These end-of-life savings may actually slow the rate of growth of health care spending, given the facts that (1) the elderly population (the large majority of those who die in any given year) is projected to grow both in absolute number and in relation to the general population, (2) the elderly use a disproportionate share of new forms of medical technology, and (3) medical technology is frequently cited as a major cause of increased medical spending.

The use of hospice care and other proposals for controlling end-of-life costs may not solve all our health care-financing problems, but with an annual saving of $30 billion -- enough to fund the District of Columbia's entire annual budget for almost seven years -- they certainly seem capable of transforming “good ethics [into] good health economics.”

John J. Mahoney
National Hospice Organization, Arlington, VA 22209

2 References

1
Emanuel EJ, Emanuel LL. The economics of dying -- the illusion of cost savings at the end of life. N Engl J Med 1994;330:540-544
Full Text | Web of Science | Medline

2
Burner ST, Waldo DR, McKusick DR. National health expenditures projections through 2030. Health Care Financ Rev 1992;14:1-29
Medline

To the Editor:

We agree with Emanuel and Emanuel that the cost savings at the end of life that are associated with advance directives may amount to only a small percentage of health care expenditures. The authors incorrectly state that we suggested that “$55 billion to $109 billion might be saved `from a policy of asking all patients about their wishes regarding life-sustaining treatment and incorporating those wishes into advance directives.' ” Our estimate was $55 billion (8.3 percent of health care expenditures that year)1. We emphasized in our article that “Our estimates are based on scanty data” and that “the data... drawn from the US Medicare population... may not accurately reflect other populations.” We viewed our estimates as a “thought experiment” and stated, “Whether this policy [of advance directives] will contribute to cost control is an empiric question that should yield to clinical research. . . . Our thought experiment could be empirically validated in a randomized clinical trial wherein the intervention is the advance directive and one of the primary outcome measures is cost of care.” After the publication of our article, Schneiderman et al.2 and Teno et al.3 reported results showing no effect of advance directives on costs, but Chambers et al.4 found a 68 percent reduction in hospital charges associated with advance directives and discussions. The final word on the effects of advance directives on health care costs is not in.

Peter A. Singer, M.D., M.P.H.
Frederick H. Lowy, M.D.
University of Toronto, Toronto, ON M5G 1L4, Canada

4 References

1
Singer PA, Lowy FH. Rationing, patient preferences, and cost of care at the end of life. Arch Intern Med 1992;152:478-480
CrossRef | Web of Science | Medline

2
Schneiderman LJ, Kronick R, Kaplan RM, Anderson JP, Langer RD. Effects of offering advance directives on medical treatments and costs. Ann Intern Med 1992;117:599-606
Web of Science | Medline

3
Teno J, Lynn J, Phillips R, et al. Do advance directives save resources? Clin Res 1993;41:551A-551A abstract.
Medline

4
Chambers CV, Diamond JJ, Perkel RL, Lasch LA. Relationship of advance directives to hospital charges in a Medicare population. Arch Intern Med 1994;154:541-547
CrossRef | Web of Science | Medline

To the Editor:

A few simple observations overwhelm the arguments of Emanuel and Emanuel.

First, nearly a third of Medicare expenditures are made in the last year of life, nearly 15 percent in the last few months, as the authors note. Modern American medical care is overwhelmingly devoted to cure and to the prolongation of life. This high expenditure represents a wasteful inability -- of physicians, patients, and society -- to come to grips with medical reality and the limitations of our technology. If we could understand this failure better, the potential for savings throughout the health care system might become clearer.

Second, it is not necessary to predict “accurately [within] months, weeks, or even days” when death will occur. That imposes a requirement of absolute foreknowledge, which we will never have. We are required to care for the sick, but not to spend huge sums to effect this caring. Rather, we must enunciate more clearly to ourselves and our society what we can and cannot do. There are many instances when, for sound medical reasons, the crash cart should be left sealed, the chemotherapeutic agents left untouched, the operating room door closed.

Finally, it is difficult to regard as “policy” decisions individual care decisions made by physicians and their colleagues working intimately with patients and families. The policy implications of the data appear more indirect; they should prompt us to eliminate misplaced requirements for care in extreme circumstances.

Simplistic solutions will not suffice. In this the Emanuels are correct.

Mark Hilberman, M.D.
8200 Kroll Way, Bakersfield, CA 93311

To the Editor:

One of the Emanuels' assumptions -- that care at the end of life is more expensive because it is labor-intensive -- is false. In hospice work, the laborers we train are the family of the hospice patient. Just as grandparents, fathers, and other family members are now trained to take care of brand-new mothers and babies (labor once provided by hospital nursery workers), so do we train the family of the hospice patient. Savings accrue from the substitution of unpaid workers for paid professionals.

Ann Mitchell, M.P.H.
Cary E. Stroud, M.D.
Greenville Hospital System, Greenville, SC 29605

To the Editor:

Senator Everett Dirksen observed that “A billion here, a billion there, and pretty soon you're talking about real money,” but Drs. Emanuel and Emanuel elevate that concept to a startling new level in their article “The Economics of Dying.” Observing that reducing the use of aggressive, life-sustaining interventions for dying patients would have saved $29.7 billion in 1993, they make the astounding conclusion that these savings are not substantial!

William D. Carroll, M.D.
719 S. Rogers, Bloomington, IN 47403-2335

Author/Editor Response

The authors reply:

To the Editor: Mitchell and Stroud claim that hospice saves money by substituting “grandparents, fathers, and other[s]... for paid professionals.” Rather than demonstrating that hospice is not labor intensive, this shows how hospices shift costs from insurers to families. This practice discriminates against women -- the predominant care givers -- as well as those who cannot afford to forgo work, and those without family nearby.

We did not intend to misquote Singer and Lowy. Like many other experts who calculate savings from the use of advance directives, however, they note that 27 percent of Medicare monies are spent on dying patients, and then assume that 27 percent of all health care spending is on dying patients1. This faulty extrapolation is the most important cause of the mistaken perception that the United States spends excessively on dying patients.

Singer and Lowy cite a recent study showing that the costs of terminal hospitalization of patients with advance directives are 68 percent lower2. In that retrospective study, the patients with advance directives were self-selected as those who wanted minimal interventions and who would obviously have minimal intervention costs. Furthermore, the focus on terminal hospitalization overstates potential savings. Available data demonstrate cost savings of 31 to 64 percent in the last 30 days of life with home hospice care; the savings decline to 15 to 27 percent when costs over a six-month period are examined. Another recent study showed that no cost savings would result from eliminating futile care3. Of the more than 10 studies on cost savings due to the use of advance directives and hospice care and the elimination of futile care, only 1 study showed substantial savings; the 2 randomized studies and the studies involving more than 1000 patients did not.

We agree with Hilberman. If Americans would accept marked limitations on the use of technology, with a concomitant reduction in the amount of testing and interventions, huge medical care cost savings could result. This change would require a metamorphosis of social values, which seems implausible given the American public's apparent unwillingness to consider rationing seriously in the current debate4.

Mahoney, Carroll, and others suggest that a saving of $29.7 billion is substantial. Our calculation was based on an idealized, best-case scenario in which all Americans completed advance directives forgoing life-sustaining treatments, accepted home hospice care, and refused futile care. This unrealizable fantasy was assumed for heuristic purposes, not to demonstrate readily attainable cost savings. Indeed, over the past 5 to 10 years there has been a reduction in the use of life-sustaining care by dying patients, a doubling in the use of hospice care by Medicare patients, and a considerable increase in the number of patients dying with do-not-resuscitate orders. Yet there have been no significant cost savings in the Medicare budget or national health spending.

Finally, several people critique our view, citing Senator Dirksen's statement “A billion here, a billion there, and pretty soon you're talking about real money.” According to Edward Schapsmeier, Dirksen's biographer, there is no documented evidence that he ever made this comment5. The last year in which he could have made such a comment would have been 1969. Before we extrapolate cost figures again without adjustment, we should realize that $1 billion of health care in 1969 would be equivalent to almost $14 billion of care in 1993 (Table 1Table 1Increase in the Gross National Product and National Medical Care Costs, 1969 to 1993.). Today, we should say “Fourteen billion here, 14 billion there, and pretty soon you're talking about real money,” and not attribute it to Senator Dirksen, who would roll over in his grave.

Ezekiel J. Emanuel, M.D., Ph.D.
Dana-Farber Cancer Institute

Linda L. Emanuel, M.D., Ph.D.
Harvard Medical School, Boston, MA 02115

5 References

1
Singer PA, Lowy FH. Rationing, patient preferences, and cost of care at the end of life. Arch Intern Med 1992;152:478-480
CrossRef | Web of Science | Medline

2
Chambers CV, Diamond JJ, Perkel RL, Lasch LA. Relationship of advance directives to hospital charges in a Medicare population. Arch Intern Med 1994;154:541-547
CrossRef | Web of Science | Medline

3
Teno J, Murphy D, Tosteson A, et al. Simulation of potential impact of a futility guideline in seriously ill adults. J Am Geriatr Soc 1994;46:A6-A6 abstract.